Influence of Severity and Types of Astigmatism on Visual Acuity in School-Aged Children in Southern China

Home , Astigmatism, Visual impairment

Int J Ophthalmol, Vol. 11, No. 8, Aug.18, 2018 www.ijo.cn Tel:8629-82245172 8629-82210956 Email:[email protected] ·Investigation· Influence of severity and types of astigmatism on visual acuity in school-aged children in southern China

Li-Li Wang, Wei Wang, Xiao-Tong Han, Ming-Guang He

Zhongshan Ophthalmic Center, State Key Laboratory of DOI:10.18240/ijo.2018.08.20 Ophthalmology, Sun Yat-Sen University, Guangzhou 510060, Guangdong Province, China Citation: Wang LL, Wang W, Han XT, He MG. Influence of severity Correspondence to: Ming-Guang He. State Key Laboratory and types of astigmatism on visual acuity in school-aged children in of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat-sen southern China. Int J Ophthalmol 2018;11(8):1377-1383 University, 54S. Xianlie Road, Guangzhou 510060, Guangdong Province, China. [email protected] INTRODUCTION Received: 2018-04-26 Accepted: 2018-05-28 stigmatism is one of the most common refractive A errors. Severe astigmatism without effective correction Abstract leads to visual impairment, amblyopia, and myopia during ● AIM: To investigate the influence of astigmatism on visual emmetropization[1-3]. Visual impairment due to astigmatism acuity in school-aged children, and to define a cutoff for may also cause insufficient educational performance and affect clinically significant astigmatism. working potential[4-6]. The prevalence of astigmatism in Chinese ● METHODS: This is a population-based, cross-sectional children ranges from 2% to 42.7%[7-11]. Previous studies also study. Among 5053 enumerated children aged 5-15y have documented that children of Hispanic and African- in Guangzhou, 3729 (73.8%) children aged 7-15 with American ethnicities had higher risk of astigmatism than non- successful cycloplegic auto-refraction (1% cyclopentolate) Hispanic white children[12]. Asian ethnicity was also related and a reliable visual acuity measurement were included. with higher risk of astigmatism[13-14]. However, these findings Ocular measurement included external eye, anterior are considerably different and not directly comparable due to segment, media and fundus and cycloplegic auto-refraction. the differences in disease definition. Primary outcome measures included the relationship Knowledge of the influence of severity and subtypes of between severity and subtypes of astigmatism and the astigmatism on visual performance is essential for a better prevalence of visual impairment. Three criteria for visual understanding of the impact of astigmatism on visual function impairment were adopted: best-corrected visual acuity and probably amblyopia. It is well established that astigmatism (BCVA) ≤0.7, uncorrected visual acuity (UCVA) ≤0.5 or <0.7 lead to visual reduction for both distant and near distance. in the right eye. However, the cutoff of clinical significant astigmatism has ● RESULTS: Increases of cylinder power was significantly not been established currently. For example, the definition associated with worse visual acuity (UCVA: β=0.051, of astigmatism varied across population-based studies. P<0.01; BCVA: β=0.025, P<0.001). A substantial increase Hashemi et al[15] used the cylinder of 0.5 D as the cutoff in the in UCVI and BCVI was seen with astigmatism of 1.00 population survey in Iran. The Sydney Myopia Study (SMS) diopter (D) or more. Astigmatism ≥1.00 D had a greater BCVI and Collaborative Longitudinal Evaluation of Ethnicity and prevalence than cylinder power less than 1.00 D (OR=4.20, Refractive Error Study (CLEERS) adopted a cylinder of 1.0 D 95%CI: 3.08-5.74), and this was also true for hyperopic, as the cutoff for astigmatism definition[12,16-17]. The Baltimore emmetropic and myopic refraction categories. Oblique Pediatric Eye Disease Study (BPEDS), Strabismus, Amblyopia astigmatism was associated with a higher risk of BCVI and Refractive Error in Singaporean Children Study (STARS) relative to with the rule astigmatism in myopic refractive and Multi-Ethnic Pediatric Eye Disease Study (MEPEDS) category (OR=12.87, 95%CI: 2.20-75.38). used a relatively high cutoff (cylinder ≥1.50 D) to examine the ● CONCLUSION: Both magnitude and subtypes of prevalence of astigmatism in preschool children[13,18-20]. astigmatism influence the prevalence of visual impairment The Refractive Error Study in Children (RESC) is a multi- in school children. Cylinder ≥1.00 D may be useful as a country large-scale population-based, cross-sectional survey cutoff for clinically significant astigmatism. of refractive error and visual impairment in school-aged ● KEYWORDS: astigmatism; visual acuity; children; children, which has used standard methodology and common population-based study definition[7,21]. Using Guangzhou RESC data, we found that 1377 Impact of astigmatism on visual acuity the overall prevalence of astigmatism (cylinder≥0.75 D) was as a minimum meridian refraction from -0.5 D to +0.75 D. 42.7% in Chinese children[7]. The objective of this study was Spherical degree was defined as the angle of the minimum to further explore the association of astigmatism with visual meridian power. Cylindrical refractive error was expressed in impairment and to identify the cutoff of astigmatism for negative cylinder form. Types of astigmatism were defined functional visual impairment. based on the minus astigmatism axis of the right eye: with SUBJECTS AND METHODS the rule (WTR; cylinder axis 0° to 30° or 150° to 180°), Study Population The RESC surveys were conducted using against the rule (ATR; cylinder axis 60° to 120°) and oblique a standardized protocol with randomized cluster sampling of (OBL; cylinder axis 31° to 59° or 121° to 149°). The visual the targeted study population, which was detailed in previous impairment was defined based on UCVA and on BCVA. Three [7,21-22] reports . The RESC survey was reviewed and approved cut-offs were used for visual impairment: UCVA≤20/40 (0.5) by human experimentation committees or Institutional Review or UCVA≤20/32 (0.7) or BCVA≤20/32 (0.7). Boards (IRB) of World Health Organization. Approval for Statistical Analysis Statistical analysis was performed using the studies at specific sites was obtained from local health Stata Version 11.0 (Stata Corporation, College Station, Texas, authorities. Written informed consent for each examined child USA). The data on the right eyes was used for analysis. The was obtained from a parent or other responsible adult during prevalence of astigmatism was categorized into 10 subgroups the household enumeration. The research followed the tenets based on cylinder severity: 0 D, 0.25 D, 0.50 D, 0.75 D, 1.00 D, of the Declaration of Helsinki. 1.25 D, 1.50 D, 1.75-2.00 D, 2.25-3.00 D and >3.00 D. Multi- Ocular Examinations Clinical examination of participants variate logistic regression analysis was used to explore the including external eye, anterior segment, media and association of visual impairment and astigmatism severity fundus were performed by trained RESC optometrists after controlling for age, sex, and spherical error. Confidence or ophthalmologist using standardized protocols. Ocular intervals (CI) and P values (significant at the <0.05 level) for alignment was assessed using the cover and uncover test at prevalence estimates and regression models were calculated near (0.5 m) and distance (4 m). Hirschberg corneal light reflex with adjustment for cluster effects associated with the sampling was used to measure the degree of tropia. design. Wald binomial confidence intervals for odds ratios Visual Acuity and Cycloplegic Refraction Uncorrected (OR) were calculated. Multivariate linear regression analysis monocular distance visual acuity was measured at 4 m with a was performed to investigate the association of astigmatism retro-illuminated logMAR chart (Precision Vision, La Salle, and logMAR visual acuity. A P-value less than 0.05 was IL, USA), in a monocular fashion with the right eye followed considered statistically significant. by the left eye, using five tumbling “E” optotypes on each line RESULTS and recorded as the smallest line read with one or no errors. If Study Population A total of 5053 children age 5-15y were uncorrected visual acuity (UCVA) in either eye was 20/40 (0.5) enumerated in Guangzhou, of whom 4364 (86.4%) children or worse, best-corrected visual acuity (BCVA) was measured were examined. Among of them 4312 (85.3%) children after cycloplegic refraction. Autorefraction was performed after cycloplegia. Cycloplegia received successful cycloplegic auto-refraction (1% was induced with two drops of 1% cyclopentolate administered cyclopentolate) and a reliable visual acuity measurement. After to each eye 5min apart and a third drop administered if a excluding 51 children with visual impairment due to non- pupillary light reflex was still present after 20min. Successful refractory causes (amblyopia, cataract, retinal disorder and cycloplegia and pupil dilation were evaluated after an other causes), 4261 children underwent further evaluation. additional 15min and was defined as a pupil dilution of more Visual acuity testing is challenging in children age 5-6y than 6 mm and without light reflex. Refractive error was because visual and cognitive function is still developing. determined by cycloplegic streak retinoscopy followed by After excluding 532 (12.49%) children age 5-6y, a total of measurement with a handheld auto-refractor (Retinomax 3729 children age 7-15y were included in the current analysis K-plus, Nikon, Tokyo, Japan) after visual acuity measurement. finally. For the current paper, auto-refraction data with complete Association of Astigmatism and Visual Impairment Figure 1 cycloplegia were used. shows the distribution of logMAR BCVA and UCVA stratified Definitions of Refractive Error and Visual Impairment by astigmatism severity of hyperopic, emmetropic and myopic Minimum meridian is the meridian with the smallest refractive spherical categories. As expected, BCVA and UCVA decreased error, the one closest to 0.0 D. Absolute minimum meridian is with a higher degree of cylinder. UCVA was more affected the absolute value of minimum meridian. Myopia was defined by astigmatism and sphere than BCVA. In children without as a minimum meridian ≤-0.75 D. Hyperopia was defined significant spherical error, mean logMAR BCVA reduced as minimum meridian ≥1.00 D. Emmetropia was defined from 0.03±0.05 to 0.06±0.07, 0.12±0.08 from no astigmatism

1378 Int J Ophthalmol, Vol. 11, No. 8, Aug.18, 2018 www.ijo.cn Tel:8629-82245172 8629-82210956 Email:[email protected]

Figure 1 Lowess smoothed cylinder-specific distribution of BCVA and UCVA of children aged 7-15y of hyperopia, emmetropia, myopia and all children A: Hyperopia is defined as sphere power (minimum meridian) +1.00 D or more; B: Emmetropia is defined as sphere power (minimum meridian) more or equal to -0.50 D and less or equal to +0.75 D; C: Myopia is defined as sphere power (minimum meridian) -0.75 D or less; D: All children.

Table 1 Prevalence of uncorrected visual impairment by severity of astigmatism Visual impairment (UCVA≤0.5) Visual impairment (UCVA≤0.7) Astigmatism (D) n Percentage (95%CI)a OR (95% CI)b n Percentage (95%CI)a OR (95% CI)b 0.00 117 21.51 (17.75-25.26) Reference 156 28.68 (23.44-33.90) Reference 0.25 323 23.01 (20.82-25.19) 1.16 (0.82-1.64) 419 29.84 (27.35-32.33) 1.19 (0.82-1.73) 0.50 205 24.12 (20.93-27.30) 1.43 (1.04-1.96) c 269 31.65 (28.43-34.86) 1.54 (1.06-2.24)c 0.75 113 29.97 (25.96-33.99) 1.52 (0.97-2.40) 140 37.14 (33.71-40.56) 1.72 (1.06-2.79)c 1.00 54 29.67 (21.70-37.64) 1.93 (0.82-4.54) 84 46.15 (37.93-54.38) 4.33 (2.06-9.11)c 1.25 29 33.72 (20.50-46.94) 1.32 (0.34-5.16) 52 60.47 (49.15-71.78) 9.61 (4.23-21.88)c 1.50 22 41.51 (24.87-58.15) 5.41 (1.92-15.20)c 35 66.04 (53.94-78.13) 14.14 (7.78-25.71)c 1.75-2.00 31 45.59 (36.43-54.74) 10.60 (5.01-22.44)c 47 69.12 (59.74-78.49) 17.94 (7.85-41.02)c 2.25-3.00 77 76.24 (68.58-83.90) 27.45 (11.34-66.43)c 86 85.15 (77.73-92.57) 23.95 (8.43-68.09)c ≥3.00 56 87.50.78.94-96.06) 107.44 (43.96-262.63)c 61 95.31 (89.30-98.43) 132.35 (33.17-528.09)c 95%CI: 95% confidence interval; D: Diopter; UCVA: Uncorrected visual acuity; OR: Odds ratio. aVisual impairment is defined as UCVA ≤0.5 or 0.7 in the right eye. bAdjusted for age, gender, and absolute minimum meridian. Binomial Wald 95%CIs were calculated. cIndicated statistical significance. to one diopter astigmatism and astigmatism ≥3.00 D; mean In the multivariate linear regression model that included UCVA was reduced to less than 0.03±0.06 to 0.09±0.12 and logMAR visual acuity, astigmatism (per diopter), age, gender, 0.38±0.13. In children with hyperopia, Mean BCVA reduced absolute minimum meridian (per diopter), statistically from 0.03±0.06 to 0.11±0.07, 0.12±0.02 from no astigmatism significant correlation was found between astigmatism to one diopter astigmatism and astigmatism >3.00 D; mean magnitude and BCVA (standardized coefficient β=0.025, UCVA was reduced to less than 0.32±0.06 to 0.13±0.08 and P<0.001), astigmatism magnitude and UCVA (standardized 0.41±0.16. In children with myopia, mean BCVA reduced from coefficient β=0.051, P<0.001). 0.04±0.07 to 0.06±0.08, 0.13±0.08 from no astigmatism to one Distribution of uncorrected visual impairment (UCVI) diopter astigmatism and astigmatism >3.00 D; mean UCVA was stratified by astigmatism was shown in Table 1 & Figure 2. reduced to less than 0.43±0.26 to 0.55±0.27 and 0.47±0.21. UCVI (UCVA≤0.5) increased from 21.51% in children without 1379 Impact of astigmatism on visual acuity

Figure 2 Proportion of visual impairment by cylinder severity in children aged 7-15y A: Proportion of UCVI (UCVA≤0.5) by cylinder; B: Proportion of UCVI (UCVA≤0.7) by cylinder; C: Proportion of BCVI (BCVA≤0.7) by cylinder.

Figure 3 Relationship between visual impairment and cylinder severity after controlling for spherical degree in children aged 7-15y A: Visual impairment (UCVA≤0.5); B: Visual impairment (UCVA≤0.7); C: Visual impairment (BCVA≤0.7). astigmatism to 87.5% of children with astigmatism ≥3.00 D Table 2 Prevalence of best-corrected visual impairment by and UCVI (UCVA≤0.7) increased from 28.68% of children severity of astigmatism without astigmatism to 95.31% of children with astigmatism Astigmatism Visual impairment (BCVA≤0.7) ≥3.00 D in the multivariate logistic regression model that (D) n Percentage (95%CI)a OR (95% CI)b included age (per year), gender, absolute minimum meridian 0.00 37 6.80 (4.40-9.20) Reference (per diopter), astigmatism (per diopter) as covariates. Children 0.25 75 5.34 (4.02-6.66) 0.79 (0. 49-1.27) with astigmatism 0.50 D, 1.00 D, more or equal to 3.00 D have 0.50 53 6.24 (4.76-7.70) 0.93 (0.58-1.49) 1.54, 4.33, 132.35 folds odds ratios for UCVI (UCVA≤0.7) 0.75 24 6.37 (3.73-9.00) 0.91 (0.53-1.57) than children without astigmatism. 1.00 27 14.84 (10.30-19.37) 2.45 (1.44-4.16)c Distribution of best-corrected visual impairment (BCVI; 1.25 20 23.26 (12.67-33.85) 4.13 (1.99-8.57)c BCVA≤0.7) stratified by astigmatism severity categories was 1.50 9 16.98 (6.97-26.99) 2.67 (1.09-6.57)c shown in Figure 2. A substantially increase of BCVI was found c 1.75-2.00 15 22.06 (9.55-34.57) 4.09 (1.86-8.99) in astigmatism of 1.00 D. BCVI raised from 6.80% of children 2.25-3.00 24 23.76 (12.30-35.23) 3.61 (1.58-8.28)c without astigmatism to 39.06% of children with astigmatism ≥3.00 25 39.06 (27.86-50.27) 8.86 (4.79-19.40)c more or equal to 3.00 D (Table 2). In the multivariate logistic 95%CI: 95% confidence interval; D: Diopter; BCVA: Best-corrected regression model that included age (per year), gender, absolute visual acuity; OR: Odds ratio. aVisual impairment is defined as minimum meridian (per diopter), astigmatism groups as BCVA≤0.7 in the right eye. bAdjusted for age, gender, and absolute covariates. Children with astigmatism 1.00 D, 1.75-2.00 D, minimum meridian. Binomial Wald 95%CIs were calculated. more or equal to 3.00 D have 2.45, 4.09, 8.86 folds times risk cIndicated statistical significance. of BCVI separately than children without astigmatism. After controlling for spherical degree, astigmatism severity has greater BCVI prevalence than the children with cylinder power considerable effect on visual impairment (Figure 3). Within less than 1.00 D (21.66%, 95%CI: 18.15%-25.17% vs 5.95%, the same spherical group, the prevalence of visual impairment 95%CI: 4.96%-6.94%). In the multivariate logistic regression increases considerably with astigmatism increased to more model that included age, gender, astigmatism≥1.00 D, absolute than 1.00 diopter. The proportion of UCVI was mainly driven minimum meridian as covariates within hyperopic, emmetropic by the sphere in myopic children but among the children with and myopic refraction categories. Proportions of BCVI were mild sphere diopters (-1 to +1 D), severe astigmatism (cylinder higher when the astigmatism ≥1.00 D and this was consistently greater than 1 D) would lead to substantial UCVI (Figure 3A found in every refraction category. (OR=10.87, 95%CI: 6.18- and 3B). Similarly, children with astigmatism ≥1.00 D had 19.06 for hyperopic refraction category; OR=7.11, 95%CI: 1380 Int J Ophthalmol, Vol. 11, No. 8, Aug.18, 2018 www.ijo.cn Tel:8629-82245172 8629-82210956 Email:[email protected] 4.46-11.34 for emmetropic refraction category; OR=2.17, 95%CI: 1.52-3.08 for myopic refraction category; OR=4.20, 95%CI: 3.08-5.74 of all children aged 7-15y). Effect of Astigmatism Subtypes on Visual Impairment Figure 4 shows the distribution of BCVI stratified by astigmatism severity and astigmatism subtypes. Proportions of BCVI were higher in OBL than WTR and ATR astigmatism when the astigmatism ≥1.00 D in emmetropic refractive category. In the multivariate logistic regression model that included age, gender, astigmatism severity, astigmatism subtypes, absolute minimum meridian as covariates within clinically significant astigmatism (astigmatism ≥1.00 D) and hyperopic, emmetropic and myopic refraction categories, OBL astigmatism were Figure 4 Influence of severity and types of astigmatism on associated with a higher risk of BCVI relative to WTR the prevalence of best-corrected visual acuity impairment astigmatism in myopic refractive category (OR=12.87, 95%CI: (BCVA≤0.7) in children aged 7-15y W: With the rule; O: Oblique; 2.20-75.38). No significant difference was found in UCVI A: Against the rule. among astigmatism subtypes. DISCUSSION corroborated with the aforementioned findings and underscores Using Guangzhou RESC data, we firstly explored the its clinical importance and highlights the need for astigmatism influence of severity and type of astigmatism on visual correction. The uncorrectable astigmatism demonstrated by acuity in the urban school children in southern China. The substantially number of people with severe astigmatism was results showed that the prevalence of visual impairment was not corrected to normal vision may indicate the association of significantly associated with both the magnitude and subtypes severe astigmatism and amblyopia. of astigmatism after controlling for age, sex, and spherical The lack of definition of clinical significant astigmatism powers. When the astigmatism ≥1.00 D, nearly 22% children contributes to the substantial discrepancies in astigmatism had BCVA≤0.7. The OR of BCVI was 10.87 times higher than prevalence across studies, which often makes it impossible non-astigmatism, indicating a cylinder power ≥1.00 D may be to analyze and compare the results directly. Various cutoffs an appropriate cutoff for clinical significant astigmatism. In of cylinder power were used for defining astigmatism. For addition, the OBL astigmatism was associated with higher risk example, 0.5 D in population-based studies in Iran and Brazil, for visual impairment than WTR at the same cylindrical levels 0.75 D in Guangzhou RESC study, 1.0 D in CLEERS and of refractive error. SMS, 1.5 D in BPEDS, STARS and MEPEDS. There is a We found that the visual acuity was nearly linear declined in clear need for an agreed definition of clinically significant distance visual acuity with increasing severity of astigmatism astigmatism, which requires more studies on how cylindrical (Figure 1). Furthermore, both the prevalence of UCVA and errors affect visual acuity. Our results based on large-scale BCVA were increased with an increased cylinder power. population showed that the prevalence of visual impairment Only few studies have reported the association between increased substantially when cylinder power of ≥1.00 D, severity of astigmatism and visual performance. Villegas indicating that this value may be an appropriate astigmatism et al[23] reported astigmatism less than 0.5 D did not affect cutoff. Astigmatism ≥1.00 D bring in a decrease in BCVA and visual acuity. Guo and Atchison[24] reported that a reduction of an increase in BCVI in hyperopic, emmetropic and myopic 0.1 logMAR line of letters was associated with an average eyes, suggesting that higher levels of cylinder produce optical 0.28 D of induced cylindrical power. Another 2 studies distortions that are not completely correctable by the usual reported that distance visual acuity declined linearly with correction of both sphere and cylinder. increasing imposed astigmatism, with 1-2 lines of logMAR The influence of stigmatism axis on visual acuity remains visual acuity per one diopter[25-26]. Other studies also have controversial. Some studies reported that reduction in distance documented that astigmatism resulted in reduction of visual acuity with induced astigmatism was greatest in ATR near visual acuity, functional visual impacts, and visual or OBL astigmatism, and least in WTR astigmatism[28-29]. developments[27]. Astigmatism of 0.50 D had clinical significant However, other studies did not observe significant correlations more risk of UCVI (UCVA≤0.7) than children without between cylinder axis and visual impairment[30-31]. The astigmatism. The current study found that every diopter discrepancy between studies may be associated with the increases of cylinder were correlated with decreases in BCVA differences in methodology, accommodation and pupil (β=0.025, P<0.001) and UCVA (β=0.051, P<0.001), which is size. Previous study reported that OBL astigmatism related

1381 Impact of astigmatism on visual acuity amblyopia require longer treatment[32]. This study observed Fundamental Research Funds of the State Key Laboratory more visual impairment due to OBL astigmatism than WTR in Ophthalmology; National Natural Science Foundation of astigmatism in the myopic refractive category, indicating that China (No.81125007); The Research Accelerator Program children with clinically significant OBL astigmatism should be was at University of Melbourne and the CERA Foundation. corrected. Further studies with large sample size and various The Centre for Eye Research Australia receives Operational ethnic origins are warranted to confirm or refute our findings. Infrastructure Support from the Victorian State Government. Astigmatism is associated with substantial social and financial Conflicts of Interest: Wang LL, None; Wang W, None; Han burden worldwide. Though the exact cost for astigmatism XT, None; He MG, None. correction was unavailable, the WHO estimated that the global REFERENCES cost of uncorrected refractory errors was 20 000 million US 1 Harvey EM. Development and treatment of astigmatism-related dollars and led to annually 202 000 million US dollar loss amblyopia. Optom Vis Sci 2009;86(6):634-639. of gross domestic product in 2007. Additional 47 000 full- 2 Gwiazda J, Grice K, Held R, McLellan J, Thorn F. Astigmatism and the time functional clinical refractionists and 18 000 ophthalmic development of myopia in children. Vision Res 2000;40(8):1019-1026. dispensers were needed[33]. For Singapore school children, 3 Twelker JD, Miller JM, Sherrill DL, Harvey EM. Astigmatism and it was reported that and annual direct cost of myopia was myopia in Tohono O’odham Native American children. Optom Vis Sci $221.68 (US$148) per subject[34]. Another study estimated 2013;90(11):1267-1273. that at least $3.8 billion annually were needed for correcting 4 Narayanasamy S, Vincent SJ, Sampson GP, Wood JM. Simulated distance vision impairment, and the majority of them were astigmatism impairs academic-related performance in children. myopia, hyperopia, and astigmatism[35]. Our results showed Ophthalmic Physiol Opt 2015;35(1):8-18. that compulsory correction of astigmatism <1.0 D may 5 Wills J, Gillett R, Eastwell E, Abraham R, Coffey K, Webber A, Wood be unnecessary because it causes low proportion of visual J. Effect of simulated astigmatic refractive error on reading performance impairment. in the young. Optom Vis Sci 2012;89(3):271-276. The strength of this study lies in the population-based design 6 Harvey EM, Miller JM, Twelker JD, Davis AL. Reading fluency and use of standardized recruitment strategy and methodology. in school-aged children with bilateral astigmatism. Optom Vis Sci This study has limitations. Firstly, the nature of cross-sectional 2016;93(2):118-125. design prevents the cause-effect inference, thus the definitive 7 He M, Zeng J, Liu Y, Xu J, Pokharel GP, Ellwein LB. Refractive evidence is needed by longitudinal data. Secondly, no near error and visual impairment in urban children in southern China. Invest visual acuity or functional impact of astigmatism was included Ophthalmol Vis Sci 2004;45(3):793-799. in the current study. Thirdly, the influence of astigmatism on 8 Xiao X, Liu WM, Ye YJ, Huang JZ, Luo WQ, Liu HT, Lin Q, Zhao visual acuity was also related to other factors, including the WX, Lin EW. Prevalence of high astigmatism in children aged 3 to 6y in higher-order aberrations, pupil size, the accommodative status, Guangxi, China. Optom Vis Sci 2014;91(4):390-396. cortex adaptation to astigmatism. Further studies adjusting 9 Chan SE, Kuo HK, Tsai CL, Wu PC. Astigmatism in Chinese primary these factors were warranted. school children: prevalence, change, and effect on myopic shift. Jpn J In conclusion, the severity and types of astigmatism Ophthalmol 2018;62(3):321-326. significantly affected both the UCVA and BCVA. Astigmatism 10 Li Z, Xu K, Wu S, Lv J, Jin D, Song Z, Wang Z, Liu P. Population- ≥1.00 D causes significantly more BCVI, suggesting that based survey of refractive error among school-aged children in rural cylinder power ≥1.00 D is an appropriate cutoff, for defining northern China: the Heilongjiang eye study. Clin Exp Ophthalmol 2014; clinically significant astigmatism. OBL astigmatism causes 42(4):379-384. more visual acuity loss than WTR at the same level of 11 Wu JF, Bi HS, Wang SM, Hu YY, Wu H, Sun W, Lu TL, Wang XR, cylindrical error. Further studies with large sample size Jonas JB. Refractive error, visual acuity and causes of vision loss in and multiple ethnicities are needed to confirm or refute our children in Shandong, China. The Shandong Children Eye Study. PLoS findings. One 2013;8(12):e82763. ACKNOWLEDGEMENTS 12 Kleinstein RN, Jones LA, Hullett S, Kwon S, Lee RJ, Friedman NE, Authors’ contributions: He MG designed the experimental Manny RE, Mutti DO, Yu JA, Zadnik K; Collaborative Longitudinal protocol and collected the data. All authors were involved in Evaluation of Ethnicity and Refractive Error Study Group. Refractive the analysis. Wang LL wrote the first draft of the manuscript. error and ethnicity in children. Arch Ophthalmol 2003;121(8):1141-1147. He MG, Wang W and Han XT reviewed and revised the 13 Fozailoff A, Tarczy-Hornoch K, Cotter S, Wen G, Lin J, Borchert M, manuscript and produced the final version. All authors read Azen S, Varma R; Writing Committee for the MEPEDS Study Group. and approved the final manuscript. Prevalence of astigmatism in 6- to 72-month-old African American Foundations: Supported by the World Health Organization and Hispanic children: the Multi-ethnic Pediatric Eye Disease Study. under National Institute of Health contract N01-EY-2103; Ophthalmology 2011;118(2):284-293.

1382 Int J Ophthalmol, Vol. 11, No. 8, Aug.18, 2018 www.ijo.cn Tel:8629-82245172 8629-82210956 Email:[email protected] 14 Huang J, Maguire MG, Ciner E, Kulp MT, Cyert LA, Quinn GE, Orel- 23 Villegas EA, Alcón E, Artal P. Minimum amount of astigmatism that Bixler D, Moore B, Ying GS; Vision in Preschoolers (VIP) Study Group. should be corrected. J Cataract Refract Surg 2014;40(1):13-19. Risk factors for astigmatism in the Vision in Preschoolers Study. Optom 24 Guo H, Atchison DA. Subjective blur limits for cylinder. Optom Vis Vis Sci 2014;91(5):514-521. Sci 2010;87(8):E549-E559. 15 Hashemi H, Nabovati P, Malekifar A, Yekta A, Ostadimoghaddam 25 Dobson V, Miller JM, Harvey EM, Mohan KM. Amblyopia in H, Jafarzadehpur E, Khabazkhoob M. Astigmatism in underserved rural astigmatic preschool children. Vision Res 2003;43(9):1081-1090. areas: a population based study. Ophthalmic Physiol Opt 2016;36(6):671-679. 26 Harvey EM, Dobson V, Miller JM. Prevalence of high astigmatism, 16 Huynh SC, Kifley A, Rose KA, Morgan I, Heller GZ, Mitchell eyeglass wear, and poor visual acuity among Native American grade P. Astigmatism and its components in 6-year-old children. Invest school children. Optom Vis Sci 2006;83(4):206-212. Ophthalmol Vis Sci 2006;47(1):55-64. 27 Read SA, Vincent SJ, Collins MJ. The visual and functional impacts 17 Huynh SC, Kifley A, Rose KA, Morgan IG, Mitchell P. Astigmatism of astigmatism and its clinical management. Ophthalmic Physiol Opt in 12-year-old Australian children: comparisons with a 6-year-old 2014;34(3):267-294. population. Invest Ophthalmol Vis Sci 2007;48(1):73-82. 28 Wolffsohn JS, Bhogal G, Shah S. Effect of uncorrected astigmatism on 18 Dirani M, Chan YH, Gazzard G, Hornbeak DM, Leo SW, Selvaraj P, vision. J Cataract Refract Surg 2011;37(3):454-460. Zhou B, Young TL, Mitchell P, Varma R, Wong TY, Saw SM. Prevalence 29 Chen SI, Hove M, McCloskey CL, Kaye SB. The effect of monocularly of refractive error in Singaporean Chinese children: the strabismus, and binocularly induced astigmatic blur on depth discrimination is amblyopia, and refractive error in young Singaporean Children (STARS) orientation dependent. Optom Vis Sci 2005;82(2):101-113. study. Invest Ophthalmol Vis Sci 2010;51(3):1348-1355. 30 Remón L, Tornel M, Furlan WD. Visual acuity in simple myopic 19 Giordano L, Friedman DS, Repka MX, Katz J, Ibironke J, Hawes astigmatism: influence of cylinder axis. Optom Vis Sci 2006;83(5): P, Tielsch JM. Prevalence of refractive error among preschool children 311-315. in an urban population: the Baltimore Pediatric Eye Disease Study. 31 Remón L, Monsoriu JA, Furlan WD. Influence of different types of Ophthalmology 2009;116(4):739-746,746.e1-e4. astigmatism on visual acuity. J Optom 2017;10(3):141-148. 20 Wen G, Tarczy-Hornoch K, McKean-Cowdin R, Cotter SA, Borchert 32 Chou YS, Tai MC, Chen PL, Lu DW, Chien KH. Impact of cylinder M, Lin J, Kim J, Varma R; Multi-Ethnic Pediatric Eye Disease Study axis on the treatment for astigmatic amblyopia. Am J Ophthalmol Group. Prevalence of myopia, hyperopia, and astigmatism in non- 2014;157(4):908-914.e1. Hispanic white and Asian children: multi-ethnic pediatric eye disease 33 Fricke TR, Holden BA, Wilson DA, Schlenther G, Naidoo KS, study. Ophthalmology 2013;120(10):2109-2116. Resnikoff S, Frick KD. Global cost of correcting vision impairment from 21 He M, Huang W, Zheng Y, Huang L, Ellwein LB. Refractive error uncorrected refractive error. Bull World Health Organ 2012;90(10):728-738. and visual impairment in school children in rural southern China. 34 Lim MC, Gazzard G, Sim EL, Tong L, Saw SM. Direct costs of Ophthalmology 2007;114(2):374-382. myopia in Singapore. Eye (Lond) 2009;23(5):1086-1089. 22 Negrel AD, Maul E, Pokharel GP, Zhao J, Ellwein LB. Refractive 35 Vitale S, Cotch MF, Sperduto R, Ellwein L. Costs of refractive Error Study in Children: sampling and measurement methods for a multi- correction of distance vision impairment in the United States, 1999-2002. country survey. Am J Ophthalmol 2000;129(4):421-426. Ophthalmology 2006;113(12):2163-2170.

1383